Studies which attempt to explain rarity within lineages are becoming increa
singly popular. An ability to explain causative factors for rarity based on
shared characteristics is an intuitively attractive concept, since it may
greatly reduce the number of studies needed to understand rarity in all spe
cies. Most studies have used the taxonomic unit of family as the lever at w
hich patterns in the distribution of rarity are examined, though no justifi
cation for the use of family has been provided. To date, no evidence has be
en produced to show that family-based patterns represent patterns at lower
taxonomic levels.
To test whether family-based levels of rarity reliably reflected generic le
vels. I compared the incidence of rarity within 114 genera from 47 families
. 113 genera (> 99%) within 46 families showed no difference from family me
ans in the proportion of rare species they contained, indicating that the i
ncidence of rarity is consistent between the taxonomic levels of family and
genus. This result suggests family level patterns reflect pattern at lower
taxonomic levels and supports the use of family as an appropriate level to
investigate patterns in the distribution of rarity within a flora.